ii6 SCIENCE PROGRESS 



will therefore be less marked there than at some distance away. 

 The following table for compressed cork illustrates this point : 



Distance from wall (mm.). Absorption coefficient (a), 

 2 -34 



5 • • • • • '38 



lo . . . , . '47 



20 . . . . . '61 



Since reverberation varies for different parts of the musical 

 scale, Taylor suggests that this result has a practical application 

 in the case of an auditorium designed for music. 



After reverberation the next most important defect in an 

 auditorium is an echo. The echo may be got rid of in two ways. 

 The form of the wall may be changed, and its surface broken 

 up by relief work, or it may be made a perfect absorber. The 

 first of these if carried out thoroughly enough to have any effect 

 is almost certain to do violence to the architectural design. 

 The second can only be accomplished by replacing the wall by 

 open windows, an impossible change. 



An important investigation on the elimination of echoes 

 was carried out by F. R. Watsori in an auditorium in the 

 University of Illinois (lo). The auditorium in question seemed 

 destined to be an acoustic horror. The reverberation was 

 first cured by Sabine's method, but echoes still made the hall 

 acoustically bad. The echoes were detected most satisfactorily 

 by means of an arc light at the focus of a parabolic reflector. 

 The beam was directed to different wall surfaces in turn and 

 the positions of the reflected beams noted. Owing to the 

 curvature of many of the surfaces there were foci of sound as 

 of light. The arc itself emitted a sound of high frequency 

 which suffered little diffraction, and the path of reflected sound 

 was traced out by the path of reflected light. 



It is, however, to be noted that the path of the beam of 

 light gives only partial information as to the path of sounds 

 of ordinary wave length. These are much more susceptible to 

 diffraction by obstacles, and have no such sharp outline as 

 beams of light ; also they require much larger surfaces for their 

 reflection. 



Watson's method of survey allowed a fairly complete diag- 

 nosis, and it was found that the echo defects were in the main 

 due to the curved wall and ceiling surfaces. 



An attempt was made to improve matters by the use of 

 plane reflecting boards placed in different positions near the 

 speaker, but these were of no avail. Considerable improve- 

 ment was effected, however, by mounting a canvas reflector, 

 whose surface was in the form of part of a paraboloid above and 

 behind the speaker so that his head was near the focus. But 



